Online Submissions: http://www.wjgnet.

com/1007-9327office World J Gastroenterol 2010 December 7; 16(45): 5669-5681

wjg@wjgnet.com ISSN 1007-9327 (print) ISSN 2219-2840 (online)
doi:10.3748/wjg.v16.i45.5669 © 2010 Baishideng. All rights reserved.

TOPIC HIGHLIGHT

Yeng S Ang, Dr., MD, Series Editor

Biomarkers in Barrett’s esophagus and esophageal

adenocarcinoma: Predictors of progression and prognosis

Chin-Ann J Ong, Pierre Lao-Sirieix, Rebecca C Fitzgerald

Chin-Ann J Ong, Pierre Lao-Sirieix, Rebecca C Fitzgerald, Key words: Barrett’s esophagus; Esophageal adenocar-
MRC Cancer Cell Unit, Hutchison-MRC Research Centre, Cam- cinoma; Esophageal dysplasia; Prognosis
bridge, CB20XZ, United Kingdom
Author contributions: Ong CAJ, Lao-Sirieix P and Fitzgerald Peer reviewers: Dr. Katerina Dvorak, Research Assistant Profes-
RC performed the literature review, critically analyzed the evi- sor, Cell Biology and Anatomy, The University of Arizona, 1501
dence and wrote the paper. N. Campbell Ave, Tucson, AZ 85724, United States; Zhiheng
Correspondence to: Rebecca C Fitzgerald, MD, MRC Can- Pei, MD, PhD, Assistant Professor, Department of Pathology and
cer Cell Unit, Hutchison-MRC Research Centre, Box 197/ Hills Medicine, New York University School of Medicine, Department
Road, Cambridge, CB20XZ, of Veterans Affairs, New York Harbor Healthcare System, 6001W,
United Kingdom. rcf@hutchison-mrc.cam.ac.uk 423 East 23rd street, New York, NY 10010, United States; Leo-
Telephone: +44-1223-763287 Fax: +44-1223-763296 nidas G Koniaris, Professor, Alan Livingstone Chair in Surgical
Received: May 27, 2010  Revised: July 28, 2010 Oncology, 3550 Sylvester Comprehensive Cancer Center (310T),
Accepted: August 4, 2010 1475 NW 12th Ave., Miami, FL 33136, United States
Published online: December 7, 2010
Ong CAJ, Lao-Sirieix P, Fitzgerald RC. Biomarkers in Bar-
rett’s esophagus and esophageal adenocarcinoma: Predictors
of progression and prognosis. World J Gastroenterol 2010;
Abstract 16(45): 5669-5681 Available from: URL: http://www.wjgnet.
com/1007-9327/full/v16/i45/5669.htm DOI: http://dx.doi.
Barrett’s esophagus is a well-known premalignant le-
org/10.3748/wjg.v16.i45.5669
sion of the lower esophagus that is characterized by
intestinal metaplasia of the squamous epithelium. It is
clinically important due to the increased risk (0.5% per
annum) of progression to esophageal adenocarcinoma
(EA), which has a poor outcome unless diagnosed early. INTRODUCTION
The current clinical management of Barrett’s esopha- Barrett’s esophagus is defined as an esophagus in which
gus is hampered by the lack of accurate predictors of the distal portion of the normal squamous lining has
progression. In addition, when patients develop EA, the been replaced by a metaplastic columnar epithelium. In
current staging modalities are limited in stratifying pa- order to make a diagnosis of Barrett’s esophagus, a seg-
tients into different prognostic groups in order to guide
ment of columnar metaplasia of any length must be vis-
the optimal therapy for an individual patient. Biomarkers
ible endoscopically above the esophagogastric junction
have the potential to improve radically the clinical man-
and be confirmed or corroborated histologically[1]. This
agement of patients with Barrett’s esophagus and EA
but have not yet entered mainstream clinical practice.
condition usually develops in the context of longstanding,
This is in contrast to other cancers like breast and pros- severe gastroesophageal reflux disease (GERD)[2], and is
tate for which biomarkers are utilized routinely to inform the only recognized precursor lesion for development of
clinical decisions. This review aims to highlight the most esophageal adenocarcinoma (EA). The incidence of EA
promising predictive and prognostic biomarkers in Bar- arising from Barrett’s esophagus is variable, depending on
rett’s esophagus and EA and to discuss what is required the grade of dysplasia associated with it. The risk of pro-
to move the field forward towards clinical application. gression to cancer increases gradually from 0.5% per year
for non-dysplastic Barrett’s, to 13% in low-grade dysplasia
© 2010 Baishideng. All rights reserved. (LGD) and 40% in high-grade dysplasia (HGD)[3,4].

WJG|www.wjgnet.com 5669 December 7, 2010|Volume 16|Issue 45|

 Ong CAJ et al . Biomarkers in Barrett’s esophagus

In Barrett’s esophagus, it is widely accepted that there no surveillance[21]. This report has suggested that surveil-
are three main histological subtypes. They include epithe- lance does not produce more quality-adjusted life years
lium that comprises mainly a gastric fundus subtype with than no surveillance, and that there is no apparent survival
parietal and chief cells; a junctional (cardia) subtype with advantage of cancer detected by surveillance due to a high
mucus-secreting glands; and the distinctive metaplastic recurrence rate and increased mortality from surgical in-
columnar epithelium with intestinal-type goblet cells[1,5]. terventions. It is hoped that biomarkers assayed in readily
These three histological subtypes occupy different zones in obtainable biological samples, such as blood or endoscop-
the esophagus. The intestinal-type metaplasia with goblet ic biopsies, can be identified to improve the clinical man-
cells is found most proximally next to the squamous epi- agement at each stage in the disease. Screening biomarkers
thelium, followed by the junctional (cardia) subtype in the could enable unidentified cases of Barrett’s esophagus to
middle, and the gastric fundus subtype most distally. The be diagnosed in the population (Figure 1, green arrow),
relevance of this subgrouping of the histological subtypes whereas predictive biomarkers could be used as adjuncts
of Barrett’s esophagus lies in the potential to develop or to replace the current surveillance program for the
malignancy. The fundic subtype has a very low risk of de- detection of dysplasia, as well as potentially being able
veloping EA malignant potential, whereas the metaplastic to predict which patients are at high risk of developing
columnar epithelium with intestinal-type goblet cells and cancer in the future (Figure 1, blue arrow). For patients
the junctional (cardia) type have a more significant risk of presenting de novo with EA, prognostic biomarkers could
malignant transformation[6,7]. This concept is important be useful to determine the best therapeutic approach and
as this together with the problem of defining Barrett’s prognosis (Figure 1, red arrow). In addition, biomarkers
esophagus based on location and length of metaplastic might have a role in determining response to treatment
epithelium has led to a detailed discussion in the American including chemopreventive agents, endoscopic treatments
Gastroenterological Association Institute technical review for patients with Barrett’s, and the use of molecular tar-
on Barrett’s esophagus. This meeting redefined Barrett’s geted therapy for those with cancer.
esophagus as “the condition in which any extent of meta-
plastic columnar epithelium that predisposes to cancer
development replaces the stratified squamous epithelium CLINICAL BIOMARKERS
that normally lines the distal esophagus”[8]. However, it is Clinical biomarkers can be defined as a characteristic that
slowly becoming apparent that the risk for development can be objectively measured or evaluated as an indicator
of EA is not solely limited to the intestinal type and that of normal biological processes, pathological processes or
better designed and powered studies are required to assess a response to a therapeutic intervention[22]. Importantly,
properly the true risk of progression in each subtype[9]. the quantification of biomarkers should aid, improve
During the development of EA, the epithelium ac- or alter clinical management. The criteria required for
cumulates multiple molecular abnormalities and becomes adoption of biomarkers into clinical use are not well de-
increasingly dysplastic[10]. The diagnosis of dysplasia al- fined. Therefore, the Early Detection Research Network
lows the progression from Barrett’s esophagus to EA to (EDRN) has defined five stages for development of bio-
be monitored by endoscopic surveillance biopsies with the markers for risk of progression[23] and similarly, McShane
aim of intervening prior to the development of invasive et al[24] have recently published recommendations for prog-
adenocarcinoma. Although randomized controlled evi- nostic tumor marker development (Figure 2). Despite the
dence is lacking, EA detected via this strategy appears to recommendation of different robust algorithms for bio-
confer a much better prognosis, as surveillance detected marker development, fewer than 12 biomarkers have been
disease is often at an early stage prior to lymph node in- approved by the US Food and Drug Administration for
volvement[11,12]. monitoring response, surveillance and recurrence of can-
There are a number of problems with this current cer at the current time[25]. This is alarming as thousands of
clinical algorithm. First of all, a significant proportion of biomarkers have been declared to be useful for diagnosis,
patients with Barrett’s esophagus are undiagnosed[13-16], surveillance or therapeutic markers for diseases. Most of
and therefore, will not benefit from any cancer prediction these biomarkers do not progress to clinical practice either
strategies. Second, surveillance is not proven to reduce due to problems developing accurate assays or because
population mortality and is based on the subjective as- the biomarker lacks sufficient sensitivity and specificity
sessment of dysplasia, which has inter and intra-observer in validation studies[26]. Clearly, a large concerted effort is
error[17-19]. Lastly, because most patients with Barrett’s needed to advance the field of biomarker discovery and
esophagus are at extremely low risk of developing EA[20], clinical implementation.
the majority are having unnecessary surveillance, which Biomarkers in Barrett’s esophagus and EA are mostly
is cumbersome both for the clinician and the patient, and selected due to their role in carcinogenesis. It is clear
poses a strain on the healthcare system. A recent review that during the transition from metaplasia to carcinoma,
to assess the cost-effectiveness of surveillance of Bar- many molecular alterations take place and they operate to-
rett’s esophagus based on a Markov model has revealed gether to influence the pathogenesis of dysplasia and EA.
that surveillance of Barrett’s esophagus for all grades of Biomarkers can be identified and investigated for their
dysplasia does more harm than good when compared to clinical applicability using two different complementary

WJG|www.wjgnet.com 5670 December 7, 2010|Volume 16|Issue 45|

 Ong CAJ et al . Biomarkers in Barrett’s esophagus

Normal squamous Barrett’s oesophagus Barrett’s oesophagus with Barrett’s oesophagus with Adenocarcinoma
low-grade dysplasia high-grade dysplasia

Population screening Predicting prognosis, best therapy and response

Predicting risk of progression and response to preventive therapy

Figure 1 Transition of squamous epithelium to intestinal metaplasia, dysplasia and adenocarcinoma, with potential useful biomarkers at each stage of the
disease. The left-most panel shows normal stratified squamous epithelium. The second panel shows Barrett’s esophagus without dysplasia, with the presence of gob-
let cells. The third and fourth panels show Barrett’s esophagus with low-grade dysplasia and high-grade dysplasia, whereas the last panel shows adenocarcinoma.

EDRN phase 1 EDRN phase 2 EDRN phase 3 EDRN phase 4 EDRN phase 5

Expression at
Decreased
Marker different stages Assay Retrospective Prospective
mortality afforded
discovery of disease development validation study validation study
by the test
progression

REMARK phase 1 EDRN phase 2 EDRN phase 3 EDRN phase 4 EDRN phase 5

Health and cost

Marker(s) associated
Marker Association with Assay External benefits from the
with survival in a
discovery survival development validation change in patient
prospective cohort
management

Figure 2 Phases of diagnostic and prognostic biomarker development proposed by the Early Detection Research Network and reporting recommenda-
tions for tumor marker prognostic studies before clinical implementation[23,24]. EDRN: Early Detection Research Network; REMARK: Reporting recommenda-
tions for tumor marker prognostic studies.

approaches[27]. The first approach is to identify candidate ers predictive of progression in Barrett’s esophagus,
biomarkers from what is currently understood about the which it is hoped could transform the current surveil-
disease process. This is a comparatively inexpensive way lance program; and (2) prognostic biomarkers in EA.
to identify putative biomarkers and possibly allow for
faster clinical implementation of the biomarker. The sec-
PROMISING BIOMARKERS IN SURVEIL-
ond method is to use a global screening approach without
an a priori hypothesis. This has become possible due to the LANCE OF BARRETT’S PATIENTS
rapid expansion of “omics” technologies, including gene Many biomarkers aimed at predicting progression in Bar-
expression analysis, epigenetics, proteomics and single rett’s patients have emerged over several years of research
nucleotide polymorphism (SNP)-based platforms. The because it is appreciated that current clinical and endo-
availability of microarray databases and other datasets on scopic criteria are unable to predict which patients are
the internet also allows for the interrogation of multiple likely to progress to EA. Biomarkers in Barrett’s esopha-
datasets to identify potential biomarkers. For example, gus can be used for population screening and early detec-
Lao-Sirieix et al[28] have identified trefoil factor 3 (TFF3) as tion of disease, confirmation of diagnosis of disease and
a promising biomarker to screen asymptomatic patients prediction of risk of progression, which determine the
for Barrett’s esophagus by comparing three publically prognosis of patients once adenocarcinoma develops and
available microarray databases. However, this approach re- predict the effectiveness of therapy. Table 1 shows a sum-
quires an intensive validation process due to the potential mary of the biomarkers that have been most extensively
for false discovery and can potentially be expensive and investigated and their potential as clinical biomarkers. In
not reproducible between laboratories. studies evaluating the efficacy of the proposed biomark-
This review focuses on two main areas: (1) biomark- ers to determine the risk of progression from Barrett’s

WJG|www.wjgnet.com 5671 December 7, 2010|Volume 16|Issue 45|

 Ong CAJ et al . Biomarkers in Barrett’s esophagus

Table 1 Summary of the most promising biomarkers for identifying patients with Barrett’s esophagus at high risk of developing
esophageal adenocarcinoma

Surveillance Highest Study Findings Statistical significance Ref.

1
biomarker EDRN stage size (n )
[29]
HGD 4 15 Progression to EA in 4 out of 15 patients with unifocal HGD RR not available
[30]
485 20 patients with HGD treated with omeprazole only developed EA RR not available
[31]
327 33 out of 76 patients with HGD developed EA RR 28 (95% CI: 13-63)
[32]
1099 12 out of 75 patients with HGD developed EA RR 12.1 (95% CI: 5-29.4)
Aneuploidy and 4 243 Panel of biomarkers (LOH of 17p and 9p and DNA abnormalities) RR 38.7 (95% CI: 10.8-138.5)
LOH (Reid Panel) can best predict progression to EA
LOH of 17p alone RR 10.6 (95% CI: 5.2-21.3) [33]
LOH of 9p alone RR 2.6 (95% CI: 1.1- 6.0)
Aneuploidy alone RR 8.5 (95% CI: 4.3-17.0)
Tetraploidy alone RR 8.8 (95% CI: 4.3-17.7)
p53 positivity by im- 3 164 Diffuse or intense TP53 staining elevated in patients who developed OR 11.7 (95% CI: 1.93-71.4) [34]
munohistochemistry EA compared to controls
48 3 out of 5 patients with low grade dysplasia who progressed to high RR not available [35]
grade dysplasia had positive p53
Mcm2 3 27 Ectopic luminal surface expression predictive of progression to HGD OR 136 (95% CI: 7.5-2464) [36]
or EA
Cyclin A 3 48 Ectopic luminal surface expression predictive of progression to HGD OR 7.6 (95% CI: 1.6-37) [37]
or EA
Methylation markers 3 53 Hypermethylation of p16 (cyclin-dependent kinase inhibitor 2A), OR 1.74 (95% CI: 1.33-2.2),
RUNX3 (Runt-related transcription factor 3) and HPP1 (transmem- 1.80 (95% CI: 1.08-2.81) and
[38]
brane protein with EGF-like and two follistatin-like domain 2) asso- 1.77 (95% CI: 1.06-2.81), re-
ciated with an increased risk of progression to high grade dysplasia spectively
or EA
195 A 8 gene methylation panel in combination with age could predict RR not available
[39]
half of progressors to HGD or EA who would not have been diag-
nosed without the use of the panel

1
Study size includes all patients in study and findings are extracted when relevant. Mcm2: Minichromosome maintenance protein 2; EA: Esophageal ad-
enocarcinoma; HGD: High-grade dysplasia; LOH: Loss of heterozygosity; EDRN: Early Detection Research Network; RR: Relative risk; OR: Odds ratio; CI:
Confidence interval.

esophagus to dysplasia and cancer, the odds ratio and thologists were misdiagnosed. Forty-two percent of these
relative risks are included whenever data were available misdiagnosed cases had only Barrett’s esophagus without
in order to give a representation of the usefulness of the dysplasia, and 8% had HGD[41]. It is clear that histological
biomarkers. differentiation between non-dysplastic Barrett’s esophagus
and LGD in particular is fraught with difficulties with poor
intra- and inter-observer agreement.
DYSPLASIA HGD is known to be a surrogate marker for the high
Dysplasia has been assessed as part of routine clinical likelihood of progression to EA. Following diagnosis of
practice for > 20 years. Although the assessment of dys- HGD, endoscopic or surgical intervention is usually con-
plasia cannot be measured objectively, it is still considered sidered. Therefore, confirmation by two independent pa-
a biomarker by most institutions, and is the current gold thologists is a pre-requisite. As a result of the practice for
standard for determining the risk for cancer progression. intervention once HGD is detected, data on progression
The current dysplasia grading system is the Vienna classifi- to EA have become much harder to obtain. Studies have
cation, which divides patients into no dysplasia, LGD and shown that the risk of progression to EA ranges from
HGD[40]. Due to its routine use, very few studies have been 16% to 59%[31,32] and a proportion of patients in whom
performed to document formally its predictive power. A HGD is detected will already harbor invasive adenocarci-
recent meta-analysis has shown that the incidence of EA noma[29,32], although with intensive biopsy protocols and
in patients undergoing surveillance for Barrett’s esophagus high definition endoscopes, this should no longer be so
rises in a stepwise manner using dysplasia as a biomarker. likely. A more ideal biomarker would be one that is less
The incidence of EA was reported to be 5.98 per 1000 subjective and that appears earlier in the pathogenetic
patient years, 16.98 per 1000 patient years and 65.8 per process, so that intervention could be considered for the
1000 patient years in Barrett’s patients without dysplasia, highest risk patients earlier in the course of their disease.
and with LGD and HGD, respectively[4]. However, histo- The evaluation of dysplasia is now well established and it
logical differentiation of the different grades of dysplasia has been suggested that other promising biomarkers are
in Barrett’s patients presents one of the most difficult more likely to be used in conjunction with the current sys-
tasks for the pathologist. In one study, 50% of Barrett’s tem than to replace the histopathological assessment of
patients who were identified to have LGD by general pa- dysplasia[42].

WJG|www.wjgnet.com 5672 December 7, 2010|Volume 16|Issue 45|

 Ong CAJ et al . Biomarkers in Barrett’s esophagus

DNA CONTENT ABNORMALITIES AND mutation of the genetic sequence, thus removing the self
repair mechanism. Reid et al[48] have performed a prospec-
LOSS OF HETEROZYGOSITY tive cohort study in 325 patients with Barrett’s esopha-
gus, and have demonstrated that LOH of chromosome
The use of DNA content abnormalities (aneuploidy and
17p(p53) significantly increased the risk of progression to
tetraploidy) and loss of heterozygosity (LOH) as bio-
cancer (relative risk of 16, 95% CI: 6.2-39). In addition,
markers to predict progression of Barrett’s esophagus to
EA has been intensively studied by the Reid group. DNA Galipeau et al[33] have demonstrated that LOH of 17p can
content abnormalities are a well-known phenomenon in be combined with LOH at 9p, DNA content abnormali-
cancer biology. A normal cell contains 46 chromosomes, ties and aneuploidy to form a panel of biomarkers to pre-
commonly referred to as 2N, and aneuploidy refers to the dict better progression of Barrett’s esophagus. This panel
state in which cells have an abnormal number of chromo- of biomarkers provides the best predictor of progression
somes. Tetraploidy, on the other hand, specifically refers to to EA to date (relative risk of 38.7, 95% CI: 10.8-138.5).
cells that have double the number of chromosomes com- Each individual marker in the panel could in itself predict
pared to normal cells (4N). In Barrett’s esophagus, numer- progression to EA with varying RR (Table 1), but when
ous studies have correlated aneuploidy and specific DNA combined together in the Reid panel, they can most accu-
abnormalities with the progression of Barrett’s esophagus rately predict progression to EA.
to EA[31,43-45], with Reid et al[31] producing the best results The panel of biomarkers that incorporate DNA ab-
by combining DNA content abnormalities with LOH. normalities and LOH, which have been developed by the
Galipeau et al[44] have demonstrated that increased 4N Reid group, are not easy to apply to the clinical setting.
(G2/tetraploid) cell populations predict progression to Efforts have therefore been made to develop alternatives.
aneuploidy, and that the development of 4N abnormali- Fang et al[45] and Vogt et al[49] have tried to circumvent the
ties is interdependent with inactivation of the p53 gene. problem of a high level technical expertise being required
Using flow cytometry and histology in a systematic endo- and the laboratory variability associated with flow cytom-
scopic biopsy protocol, Reid et al[31] first described the use etry, by using image cytometric DNA analysis in smaller
of aneuploidy and increased 4N fractions as biomarkers studies. In these studies, they have concluded that image
to identify subsets of patients with Barrett’s esophagus cytometry can provide a more sensitive marker than us-
at low and high risk of developing EA. Using a cut-off ing HGD to identify groups of patients who are likely to
for 4N fractions of > 6% as abnormal, Reid has reported progress to EA, and have highlighted that image cytome-
that the relative risk of cancer for these patients compared try has significant advantage over flow cytometry in terms
to those below this cut-off value was 7.5 (95% CI: 4-14). of costs and practicality. These findings, while promis-
In addition, patients who had baseline aneuploidy had a ing, still require validation with a much larger sample
relative risk of cancer of 5 (95% CI: 2.7-9.4) compared to size. The development of high-fidelity DNA histograms
patients who did not have baseline aneuploidy. generated by automated software to measure aneuploidy
p16 and p53 are two commonly studied tumor sup- further strengthens the role of DNA abnormalities as a
pressor genes that reside on chromosome 9p and 17p, biomarker to predict progression in Barrett’s patients[50,51].
respectively. These two tumor suppressor genes can be Other interesting novel techniques to measure aneuploidy
silenced via LOH, mutations and DNA methylation. Si- and other chromosomal aberrations have also been de-
lencing of the p16 allele is thought to be one of the earli- scribed in the literature. Li et al[52] have demonstrated that
est events in Barrett’s esophagus, which results in clonal the number of SNPs was highly correlated with chromo-
expansion[46]. However, a recent study by Leedham et al[47] somal abnormalities in Barrett’s esophagus and EA, and
has demonstrated that Barrett’s esophagus can arise from have suggested that SNP-based genotyping could possibly
multiple independent clones, which results in clonal het- be used to stratify the cancer risk in patients with Barrett’s
erogeneity. This study was performed by investigating esophagus.
individual crypts microdissected from esophagectomy As mentioned previously, the use LOH as biomarkers
specimens that contained adenocarcinoma and associated is not without its own problems. The detection of LOH is
dysplasia, to detect clonal heterogeneity not detected by complex and requires the collection of snap frozen sam-
whole biopsy analysis. Overall, p16 by itself is unlikely to ples, followed by extraction of DNA and an amplification
be an ideal biomarker to predict progression because it ap- step prior to polymerase chain reaction analysis[53]. This is
pears too early in the pathogenesis, and it has been shown in addition to the high costs needed to build and maintain
that there is no evidence of association between silencing facilities to enable the use of this panel of biomarkers in
of p16 and grade of dysplasia[46]. p53 LOH, on the other routine medical institutions. An alternative method would
hand, provides one of the most promising biomarkers to be to use fluorescence in situ hybridization (FISH) to de-
predict progression of Barrett’s esophagus, as part of the tect LOH, but this method is limited by poor sensitivity
Reid panel. p53 is a nuclear tumor suppressor protein that (68.4%) when compared to genotyping[54].
is responsible for the integrity of the genetic sequence. Immunostaining for p53 provides another alternative
Any damage to DNA should result in increased expres- to genotyping of chromosome 17p to predict progression
sion of p53, which causes cells to arrest at the G1 phase of Barrett’s esophagus because the presence of p53 muta-
to allow for DNA repair, and if this is not possible, then tions can often cause protein accumulation, which allows
apoptosis ensues. Silencing of p53 can occur via LOH or for detection by immunohistochemistry[34,35]. Although

WJG|www.wjgnet.com 5673 December 7, 2010|Volume 16|Issue 45|

 Ong CAJ et al . Biomarkers in Barrett’s esophagus

the use of immunostaining of p53 allows easy clinical patients with biopsies that express cyclin A at the surface
implementation, its efficacy as a biomarker is limited, and were more likely to progress to EA than those who did
positive staining was only seen in one third of patients not (OR: 7.5, 95% CI: 1.8-30.7)[37]. Prospective studies are
in a nested case-control study to evaluate the efficacy of required to determine properly the usefulness of cyclins
immunostaining for p53 as a marker to predict progres- as predictive biomarkers.
sion[34]. This is because staining for p53 does not always
correlate with mutations. In instances in which mutations
result in deletion or truncation of p53, it will not be de- EPIGENETIC CHANGES
tected by immunostaining. Epigenetic changes (or non-DNA sequence changes) in the
In summary, the detection of aneuploidy and DNA form of hypomethylation, hypermethylation and alteration
content abnormalities in the Reid panel appears to be one to histone complexes have also been found to be implicat-
of the most promising biomarker panels to detect the ed in the pathogenesis of Barrett’s esophagus and EA[38,59].
progression of Barrett’s esophagus to EA. However, tech- Hypermethylation of promoter Cpg island is thought to be
nical difficulties that have hindered the use of analysis of the cause of transcriptional silencing of tumor suppressor
DNA content abnormalities in the Reid panel need to be genes such as CDKN2A (p16), APC, CDH1 (E-cadherin),
addressed. SNP analysis or image cytometry are other al- and ESR1 (ER, estrogen receptor α)[59]. Hypermethylation
ternative techniques used to measure aneuploidy and other of these genes is usually found in a large contiguous field,
chromosomal aberrations but remains to be validated in which suggests possible clonal expansion of hypermethyl-
larger studies. ated cells or hypermethylation of a field of metaplastic
cells[59]. Further work on the methylation status of promot-
er regions of genes has revealed that methylation of p16
PROLIFERATION MARKERS (OR: 1.74, 95% CI: 1.33-2.20), RUNX3 (OR: 1.80, 95%
Dysplasia is typically described as being associated with CI: 1.08-2.81) and HPP1 (OR: 1.77, 95% CI: 1.06-2.81) in
abnormal cellular proliferation and differentiation[55,56]. patients with non-dysplastic Barrett’s esophagus and LGD
Our laboratory and others have demonstrated abnormal were independent risk factors for progression to HGD
surface staining of markers of proliferation [minichro- and EA[38]. More recently, Jin et al[60] have demonstrated
mosome maintenance protein (Mcm) 2, 5 and Ki67] in that a methylation biomarker panel that comprises eight
dysplastic Barrett’s mucosa[36,55,56]. This finding has served genes could accurately determine the risk of progression
as the basis for the use of aberrant surface expression in patients with Barrett’s esophagus in a retrospective, mul-
of Mcm2, together with a brushing technique to predict ticenter validation study. In that study, promoter methyla-
progression in patients with Barrett’s esophagus[36]. How- tion levels of eight genes were quantified by methylation-
ever, large prospective studies are needed before they specific PCR in patients who did not progress (n = 145)
can be used in routine clinical practice. compared to those who did progress (n = 50) to HGD
or EA. Receiver operating characteristics curves were
constructed to evaluate the usefulness of the eight-gene
CELL CYCLE MARKERS methylation panel and the authors have concluded that,
Members of the cyclin family such as cyclin A and D are with specificity set at 0.9, the eight-gene methylation panel
also interesting biomarkers for Barrett’s esophagus. Cy- in combination with age predicted half the progressors
clin D is a proto-oncogene protein and overexpression in who would not have been diagnosed without using these
Barrett’s esophagus results in inappropriate phosphoryla- biomarkers. Similarly, a recent study by Wang et al[61] has
tion and inactivation of p105-Rb. Increased expression shown that hypermethylation of p16 and APC was a good
of cyclin D has been implicated in the predisposition to predictor of progression to HGD or EA [OR: 14.97, 95%
transform from metaplastic epithelium to cancer. and can CI (1.73, inf)]. The fact that methylation changes in DNA
possibly be a useful biomarker in identifying patients with occur early in the progression from Barrett’s esophagus
Barrett’s esophagus at high risk of developing EA[57,58]. to dysplasia suggest that they could potentially be used as
Bani-Hani et al[58] have performed a case-control study and biomarkers to predict which groups of patients are likely
have shown that Barrett’s patients who are positive for to progress to dysplasia and EA[59,62]. However, the main
cyclin D detected via immunohistochemistry were more problem of the utility of hypermethylation as biomarkers
likely to develop EA (OR: 6.85, 95% CI: 1.57-29.91). lies in the fact that techniques that have been applied for
These findings were however not replicated in a larger detection of epigenetic changes require enzyme digestion,
population-based case-control study performed by Murray affinity enrichment or bisulfite treatment before probe
et al[34]. In that study, only immunohistochemical detection hybridization or sequencing can be done to detect meth-
of p53 has been shown to be a useful biomarker for ma- ylation in samples. These arrays of techniques are far too
lignant progression in Barrett’s esophagus. Cyclin A is ex- technically demanding and time consuming for routine
pressed just before the beginning of DNA synthesis and utilization in the clinic[63-72].
is an important check mechanism in the G1-S transition
of the cell cycle. In a case-control study, surface expres-
sion of cyclin A in Barrett’s esophagus samples has been PROGNOSTIC BIOMARKERS IN EA
shown to be correlated with the degree of dysplasia, and The overall 5-year survival for EA remains < 14%[73].

WJG|www.wjgnet.com 5674 December 7, 2010|Volume 16|Issue 45|

 Ong CAJ et al . Biomarkers in Barrett’s esophagus

The current staging of EA is the internationally recog- survival, lymph node involvement or response to chemo-
nized TNM system[74], which is based exclusively on the therapy. The advantage of using these methods is that
anatomical extent of the disease. This is assessed using they allow the hypothesis-free interrogation of many
a combination of tumor depth (T), number of lymph targets simultaneously. Table 3 gives a summary of the
nodes involved (N), and presence or absence of metasta- molecular signatures discovered by microarray technol-
sis (M). The TNM system remains useful for staging of ogy, including the methodology used. However, despite
esophageal tumors because patients with more advanced the number of studies, none of these molecular signa-
stage disease clearly do worse than those in the early tures or techniques to stratify patients with EA has yet
stage of the disease. For patients deemed to have poten- reached clinical utility. This is in contrast to other cancers
tially curative disease (T3N1 or less), surgical treatment for which prognostic signatures are starting to be used in
with or without chemotherapy provides the only chance the clinical setting[101-103]. In EA, molecular signatures have
of cure, but it is highly invasive and has a high morbidity usually been generated from underpowered cohorts and
rate. Biomarkers that can accurately predict the progno- many studies have combined molecular profiling of both
sis of patients with this disease could aid in the selection EA and squamous cell carcinoma of the esophagus in
of patients most likely to benefit from surgery. In addi- the same study. It is known that the molecular profile of
tion, it is also hoped that biomarkers can identify differ- squamous cell carcinoma and EA is different[104,105], and
ent subgroups of tumors that will benefit from specific for accurate prognosis, studies should differentiate be-
treatment, including molecularly targeted treatments. tween these two types of tumors. An additional problem,
Prognostic biomarkers in patient with EA have com- which is not dissimilar to biomarkers discovered for the
monly been studied to determine the association with the transition of Barrett’s esophagus to EA, is that the tech-
following outcome and tumor characteristics: (1) survival; nique used might not be applicable to routine laboratories
(2) lymphovascular invasion and metastasis; and (3) re- and will therefore be expensive. It is therefore important
sponse to chemotherapy and radiotherapy. that researchers also consider how best to apply molecu-
Traditional candidate approaches for analyzing gene lar biomarkers to the clinic, and they should consider
and protein expression in cancer have identified a large validation using methods such as immunohistochemis-
number of biomarkers that have important prognostic try. Whichever method is used, data reproducibility and
value. These biomarkers can be considered in terms of validation in independent samples are perhaps the most
the six classical hallmarks described by Hanahan et al[75], important factors to determine whether molecular signa-
with inflammation added as the seventh hallmark recent- tures are adopted for clinical application. This problem is
ly. They include: (1) self sufficiency in growth signals; (2) becoming increasingly recognized, and many reviews have
insensitivity to growth inhibitory (antigrowth) signals; (3) reiterated the need for validation of molecular signatures
evasion of programmed cell death (apoptosis); (4) limit- and the development of assays that have general clinical
less replicative potential; (5) sustained angiogenesis; (6) applicability[106-112].
invasion and metastasis; and (7) cancer-related inflamma-
tion.
Table 2 gives an overview of the biomarkers in each CONCLUSION
category and their association with survival or surrogate The pathogenesis from Barrett’s esophagus to EA is high-
measures of prognosis. This list is not exhaustive but it ly complex. Multiple molecular alterations occur during
highlights the important biomarkers that have been in- this process, which leads to a heterogenous tumor by the
vestigated and reported to be prognostic. A recent review time that EA develops. Biomarkers can complement the
by Lagarde et al[76] has described in greater detail many current clinical management of Barrett’s esophagus and its
of these biomarkers and their molecular basis. It is well transition to EA in three main ways. They can be used to:
known that many of these molecular alterations occur in identify patients not previously diagnosed with Barrett’s
tandem during the progression of Barrett’s esophagus to esophagus via population screening; improve the surveil-
EA and are present to varying degrees. These biomarkers lance of patients with Barrett’s esophagus; and identify
have been shown to be associated with survival or tumor prognostic groups and best therapy once EA develops.
characteristics, but subsequent replication of findings, as There has already been a tremendous amount of
required for the EDRN validation of biomarkers, is of- research done to create an ideal biomarker or panel of
ten lacking. It is highly unlikely that any of these markers biomarkers to predict accurately progression of Barrett’s
by itself can predict survival accurately because several esophagus to dysplasia or EA. This is in conjunction
molecular alterations can operate together to influence with large amounts of resources and money spent in
the pathogenesis of EA. Again, generating panels of bio- laboratories and in clinical trials as the research is be-
markers to create a molecular signature in EA could be ing conducted. Although no biomarkers have been able
useful in determining the prognosis of patients with EA. to replace the current gold standard of dysplasia as a
biomarker in routine clinical practice, it is reassuring
to know that certain biomarkers hold great promise to
MOLECULAR SIGNATURE OF EA transit from the bench to the bedside. It is becoming
Several studies have used microarray technologies to increasingly clear that one biomarker by itself is highly
generate molecular signatures that correlate with overall unlikely to predict progression with high sensitivity and

WJG|www.wjgnet.com 5675 December 7, 2010|Volume 16|Issue 45|

 Ong CAJ et al . Biomarkers in Barrett’s esophagus

Table 2 Summary of the biomarkers and the prognostic impact in esophageal adenocarcinoma

Category of cell Biomarker Sample Endpoint Findings Statistical Ref.

alteration size (n ) significance
[77]
Self sufficiency in Cyclin D 124 Survival 2 of 3 genotypes confers a poorer overall survival P = 0.0003
growth signals EGFR 103 Survival Expression showed a trend towards a correlation with poorer P = 0.07 [78]
overall survival
75 Survival Decreased expression correlated with poorer survival on uni- P = 0.034 [79]
variate analysis only
[80]
Ki-67 59 Survival Low levels (< 10%) of staining correlated with poorer survival HR: 3.9, P = 0.02
[81,82]
Her2/neu 63 Survival Amplification detected by FISH correlated with poorer survival P = 0.03
[83]
TGF-a 61 Survival Low levels significantly correlated with cancer specific death P = 0.03
87 Tumor progression, High levels significantly correlated with:
lymph node metastasis Tumor progression P = 0.025 [84]

Lymph node metastasis P < 0.05

Insensitivity to TGF-b1 123 Survival Overexpression correlated with poorer survival on univariate P = 0.0255 [85]
growth inhibi- analysis only
tory (antigrowth) 57 Survival High plasma levels correlated with poorer overall survival P = 0.0317 [86]

signals APC 52 Survival High plasma levels of methylation of APC associated with P = 0.016 [87]
poorer survival
P21 30 Survival Alteration in expression after chemotherapy correlated with P = 0.011 [88,89]
better survival
Evasion of P53 30 Survival Alteration in expression after chemotherapy correlated with P = 0.011 [88]
programmed cell better survival
death (apoptosis) Bcl-2 35 Survival Expression correlated with poorer survival P = 0.03 [90]

COX-2 100 T-stage, N-stage, tumor Higher levels expression correlated with:
recurrence and survival Higher T-stage, P = 0.008
[91]
Higher N-stage, P = 0.049
Increased risk of tumor recurrence P = 0.01
Poor survival P < 0.001
[92]
20 Survival Strong staining correlated with poorer survival P = 0.03
145 Distant metastasis, local Strong staining correlated with:
recurrence and survival Distant metastasis P = 0.02 [93]

Local recurrence P = 0.05

Poorer survival P = 0.002
NF-kB 43 Survival Activated NF-κB predictive of:
[94]
Poorer disease free survival P = 0.010
Poorer overall survival P = 0.015
Limitless replica- Telomer- 46 Survival Higher telomere-length ratio shown to be an independent poor P < 0.02 [95]
tive potential ase prognostic factor
Sustained angio- CD105 75 Survival, angiolym- Significant correlation between expression and:
genesis phatic invasion, lymph Poorer survival P < 0.01
node metastasis and Presence of angiolymphatic invasion P < 0.05 [96]
tumor stage and distant More lymph node metastasis P < 0.01
metastasis Higher tumor stage P < 0.001
More distant metastasis P < 0.01
VEGF 75 Survival, angiolym- Significant correlation between high expression and:
phatic invasion, lymph Poorer survival P < 0.01
node metastasis, stage Presence of angiolymphatic invasion P < 0.05 [96]
of tumor and distant More lymph node metastasis P < 0.01
metastasis Higher stage of tumor P < 0.01
More distant metastasis P < 0.01
[80]
Tissue invasion Cadherin 59 Survival Reduced level correlated with poorer overall survival HR: 3.3, P = 0.05
and metastasis uPA 54 Survival High uPA correlated with poorer survival P = 0.0002 [97]

TIMP 24 Survival and disease Reduction of expression correlated with poorer overall survival P = 0.007 [98]
stage and higher disease stage P = 0.046
Others Promoter 41 Survival and tumor Earlier tumor recurrence and poorer overall survival if > 50% of P = 0.05 [99]
hyper- recurrence gene profile methylated P = 0.04
methyl- 84 Differentiation Hypermethylation of MGMT (Methylated-DNA-protein-cysteine P = 0.0079
ation methyltransferase) gene correlated with: [100]

Higher tumor differentiation

EGFR: Epidermal growth factor receptor; Her2/neu: Human EGFR2; TGF: Transforming growth factor; APC: Adenomatosis polyposis coli; P21: Cyclin-
dependent kinase inhibitor 1; Bcl-2: B-cell lymphoma 2; COX-2: Cyclooxygenase-2; NF-κB: Nuclear factor-κB; CD105: Endoglin; VEGF: Vascular endothelial
growth factor; uPA: Urokinase-type plasminogen activator; TIMP: Tissue inhibitor of metalloproteinase.

specificity. Panels of biomarkers such as the eight-gene predict progression, seem to provide the most accurate
methylation panel or the Reid panel, which combine predictor of progression based on statistics. Unfortu-
LOH at various loci and DNA content abnormalities to nately, the common theme in these panels of markers is

WJG|www.wjgnet.com 5676 December 7, 2010|Volume 16|Issue 45|

 Ong CAJ et al . Biomarkers in Barrett’s esophagus

Table 3 Summary of the molecular signatures discovered by microarray technology and latest methods used to correlate molecular
alterations and prognosis in patients with esophageal adenocarcinoma

Method Sample Outcome Findings Statistical significance External Ref.

size (n ) validation
[113]
Oligonucleotide 75 Survival A 4-gene signature prognosticated patients P = 0.0001 Yes
cRNA microarray 77 Lymphatic Created a gene signature predicting lymph node Argininosuccinate synthetase No [114]
spread metastasis expression (ASS) (P = 0.048)
19 Chemotherapy Unsupervised hierarchical clustering divided Not statistically significant No
[115,116]
response patients into 2 groups, one of which responded to
preoperative chemotherapy
47 Chemotherapy 86 genes dysregulated P < 0.001 No
[117]
response Ephrin B3 expression associated with chemotherapy
response, tumor grading and stage
Oligonucleotide 46 Chemotherapy Gene signature not predictive in adenocarcinoma of Not statistically significant No [118]
cDNA microarray response esophagus
Proteomic analysis 34 Chemotherapy HSP27 expression associated with response to che- P < 0.05 No [119]
response motherapy
Single nucleotide 210 Survival and 5 polymorphisms in 3 genes associated with longer P = 0.004 No [120]
polymorphism recurrence recurrence free survival and reduced recurrence
[121]
microRNAs analysis 96 Survival Low miR-375 levels associated with worse survival P = 0.002 No
Multiplex ligation- 33 Survival Patients with more than 12 chromosomal aberra- P = 0.014 No
[122]
dependent probe tions had a poorer outcome than patients with < 12
amplification

that they are far too expensive to be applied in routine has a slightly different focus, much could be gained these
clinical use, and technical expertise is not available in all collaborative efforts if a proportion of the resources and
centers to utilize these panels of biomarkers. The issue patient samples could be used to validate biomarkers in
of costs and practicality of biomarkers should be one Barrett’s or tumor samples.
of the principle considerations before research and re- Lastly, biomarkers should be seen as adjuncts to aid
sources are channeled into it. clinical management of patients with Barrett’s esophagus
Although traditional methods of identifying biomark- and EA rather than in isolation in predicting the risk of
ers in Barrett’s esophagus and its transition to dysplasia progression, prognosis or response to therapy. As such,
and EA have helped greatly in the understanding of the clinical factors in conjunction with biomarkers should be
disease process, new technologies to create molecular incorporated into a model that can accurately determine
signatures have also helped by identifying many impor- the desired outcome. Such models have been used in oth-
tant biomarkers not previously thought to be involved in er cancers and diseases such as the MELD score for liver
its pathogenesis. A few biomarkers identified from both disease or the Nottingham prognostic index for breast
traditional methods and new technological platforms have cancer. Upon generation and validation of the model, it
shown great potential in predicting the progression from should then be rigorously validated in an independent
Barrett’s esophagus to EA. However, a concerted effort large cohort of patients in a prospective fashion. In fu-
is still needed to validate these biomarkers or molecular ture, patients can then be risk stratified based on a score
signatures in independent, large-scale prospective cohorts to determine the treatment strategy, hence individualizing
and to develop inexpensive, practical assays to allow for treatment to improve patient care and outcome.
clinical applicability. Realistically, this can only be achieved
by a multicenter collaboration to tackle the challenges of
ACKNOWLEDGMENTS
the large amount of resources, scientific and clinical input
required to advance the field of biomarkers in Barrett’s We would like to thank Dr. Elizabeth Bird-Lieberman for
esophagus. There are a few major collaborations in the her kind contributions of the scanned histology slides for
United Kingdom to date, and they include the Chemopre- Figure 1.
vention of Premalignant Intestinal neoplasia trial (CHO-
PIN) and Oesophageal Cancer Clinical and Molecular REFERENCES
Stratification Study (OCCAMS). This is also mirrored in
1 Playford RJ. New British Society of Gastroenterology (BSG)
the international arena with Barrett’s Esophagus and Ade-
guidelines for the diagnosis and management of Barrett's
nocarcinoma Consortium (BEACON) and Asian Barrett’s oesophagus. Gut 2006; 55: 442
Consortium as two examples of collaborative work on 2 Ronkainen J, Aro P, Storskrubb T, Johansson SE, Lind T,
Barrett’s esophagus. These initiatives allow for the pooling Bolling-Sternevald E, Vieth M, Stolte M, Talley NJ, Agréus L.
of resources, expertise and knowledge between centers Prevalence of Barrett’s esophagus in the general population:
an endoscopic study. Gastroenterology 2005; 129: 1825-1831
and allow for the recruitment of large numbers of pa- 3 Curvers WL, ten Kate FJ, Krishnadath KK, Visser M, Elzer B,
tients that are necessary to advance the field of biomark- Baak LC, Bohmer C, Mallant-Hent RC, van Oijen A, Naber
ers in Barrett’s esophagus and EA. Although each study AH, Scholten P, Busch OR, Blaauwgeers HG, Meijer GA,

WJG|www.wjgnet.com 5677 December 7, 2010|Volume 16|Issue 45|

 Ong CAJ et al . Biomarkers in Barrett’s esophagus

Bergman JJ. Low-grade dysplasia in Barrett’s esophagus: Thornquist M, Winget M, Yasui Y. Phases of biomarker
overdiagnosed and underestimated. Am J Gastroenterol 2010; development for early detection of cancer. J Natl Cancer Inst
105: 1523-1530 2001; 93: 1054-1061
4 Wani S, Puli SR, Shaheen NJ, Westhoff B, Slehria S, Bansal A, 24 McShane LM, Altman DG, Sauerbrei W, Taube SE, Gion
Rastogi A, Sayana H, Sharma P. Esophageal adenocarcinoma M, Clark GM. REporting recommendations for tumour
in Barrett’s esophagus after endoscopic ablative therapy: a MARKer prognostic studies (REMARK). Br J Cancer 2005;
meta-analysis and systematic review. Am J Gastroenterol 2009; 93: 387-391
104: 502-513 25 Anderson NL, Anderson NG. The human plasma proteome:
5 Paull A, Trier JS, Dalton MD, Camp RC, Loeb P, Goyal RK. history, character, and diagnostic prospects. Mol Cell Pro-
The histologic spectrum of Barrett’s esophagus. N Engl J teomics 2002; 1: 845-867
Med 1976; 295: 476-480 26 Srivastava S, Verma M, Henson DE. Biomarkers for early
6 Spechler SJ, Goyal RK. The columnar-lined esophagus, detection of colon cancer. Clin Cancer Res 2001; 7: 1118-1126
intestinal metaplasia, and Norman Barrett. Gastroenterology 27 Mitra AP, Bartsch CC, Cote RJ. Strategies for molecular ex-
1996; 110: 614-621 pression profiling in bladder cancer. Cancer Metastasis Rev
7 Kelty CJ, Gough MD, Van Wyk Q, Stephenson TJ, Ackroyd 2009; 28: 317-326
R. Barrett’s oesophagus: intestinal metaplasia is not essen- 28 Lao-Sirieix P, Boussioutas A, Kadri SR, O’Donovan M,
tial for cancer risk. Scand J Gastroenterol 2007; 42: 1271-1274 Debiram I, Das M, Harihar L, Fitzgerald RC. Non-endo-
8 Wang KK, Sampliner RE. Updated guidelines 2008 for the scopic screening biomarkers for Barrett’s oesophagus: from
diagnosis, surveillance and therapy of Barrett’s esophagus. microarray analysis to the clinic. Gut 2009; 58: 1451-1459
Am J Gastroenterol 2008; 103: 788-797 29 Weston AP, Sharma P, Topalovski M, Richards R, Cherian
9 Riddell RH, Odze RD. Definition of Barrett’s esophagus: R, Dixon A. Long-term follow-up of Barrett’s high-grade
time for a rethink--is intestinal metaplasia dead? Am J Gas- dysplasia. Am J Gastroenterol 2000; 95: 1888-1893
troenterol 2009; 104: 2588-2594 30 Overholt BF, Lightdale CJ, Wang KK, Canto MI, Burdick S,
10 Fitzgerald RC. Molecular basis of Barrett's oesophagus and Haggitt RC, Bronner MP, Taylor SL, Grace MG, Depot M.
oesophageal adenocarcinoma. Gut 2006; 55: 1810-1820 Photodynamic therapy with porfimer sodium for ablation of
11 van Sandick JW, van Lanschot JJ, Kuiken BW, Tytgat GN, high-grade dysplasia in Barrett’s esophagus: international,
Offerhaus GJ, Obertop H. Impact of endoscopic biopsy sur- partially blinded, randomized phase III trial. Gastrointest
veillance of Barrett’s oesophagus on pathological stage and Endosc 2005; 62: 488-498
clinical outcome of Barrett’s carcinoma. Gut 1998; 43: 216-222 31 Reid BJ, Levine DS, Longton G, Blount PL, Rabinovitch
12 Wong T, Tian J, Nagar AB. Barrett’s surveillance identifies PS. Predictors of progression to cancer in Barrett’s esopha-
patients with early esophageal adenocarcinoma. Am J Med gus: baseline histology and flow cytometry identify low-
2010; 123: 462-467 and high-risk patient subsets. Am J Gastroenterol 2000; 95:
13 Schlansky B, Dimarino AJ Jr, Loren D, Infantolino A, Kow- 1669-1676
alski T, Cohen S. A survey of oesophageal cancer: pathol- 32 Schnell TG, Sontag SJ, Chejfec G, Aranha G, Metz A, O’
ogy, stage and clinical presentation. Aliment Pharmacol Ther Connell S, Seidel UJ, Sonnenberg A. Long-term nonsurgical
2006; 23: 587-593 management of Barrett’s esophagus with high-grade dys-
14 Gerson LB, Shetler K, Triadafilopoulos G. Prevalence of plasia. Gastroenterology 2001; 120: 1607-1619
Barrett’s esophagus in asymptomatic individuals. Gastroen- 33 Galipeau PC, Li X, Blount PL, Maley CC, Sanchez CA, Odze
terology 2002; 123: 461-467 RD, Ayub K, Rabinovitch PS, Vaughan TL, Reid BJ. NSAIDs
15 Rex DK, Cummings OW, Shaw M, Cumings MD, Wong RK, modulate CDKN2A, TP53, and DNA content risk for pro-
Vasudeva RS, Dunne D, Rahmani EY, Helper DJ. Screening gression to esophageal adenocarcinoma. PLoS Med 2007; 4:
for Barrett’s esophagus in colonoscopy patients with and e67
without heartburn. Gastroenterology 2003; 125: 1670-1677 34 Murray L, Sedo A, Scott M, McManus D, Sloan JM, Hardie
16 Ward EM, Wolfsen HC, Achem SR, Loeb DS, Krishna M, LJ, Forman D, Wild CP. TP53 and progression from Barrett’
Hemminger LL, DeVault KR. Barrett’s esophagus is com- s metaplasia to oesophageal adenocarcinoma in a UK popu-
mon in older men and women undergoing screening colo- lation cohort. Gut 2006; 55: 1390-1397
noscopy regardless of reflux symptoms. Am J Gastroenterol 35 Weston AP, Banerjee SK, Sharma P, Tran TM, Richards R,
2006; 101: 12-17 Cherian R. p53 protein overexpression in low grade dys-
17 Reid BJ, Haggitt RC, Rubin CE, Roth G, Surawicz CM, Van plasia (LGD) in Barrett’s esophagus: immunohistochemical
Belle G, Lewin K, Weinstein WM, Antonioli DA, Goldman H. marker predictive of progression. Am J Gastroenterol 2001;
Observer variation in the diagnosis of dysplasia in Barrett’s 96: 1355-1362
esophagus. Hum Pathol 1988; 19: 166-178 36 Sirieix PS, O’Donovan M, Brown J, Save V, Coleman N,
18 Montgomery E, Bronner MP, Goldblum JR, Greenson JK, Fitzgerald RC. Surface expression of minichromosome
Haber MM, Hart J, Lamps LW, Lauwers GY, Lazenby AJ, maintenance proteins provides a novel method for detecting
Lewin DN, Robert ME, Toledano AY, Shyr Y, Washington patients at risk for developing adenocarcinoma in Barrett’s
K. Reproducibility of the diagnosis of dysplasia in Barrett esophagus. Clin Cancer Res 2003; 9: 2560-2566
esophagus: a reaffirmation. Hum Pathol 2001; 32: 368-378 37 Lao-Sirieix P, Lovat L, Fitzgerald RC. Cyclin A immuno-
19 Reid BJ, Li X, Galipeau PC, Vaughan TL. Barrett’s oesopha- cytology as a risk stratification tool for Barrett’s esophagus
gus and oesophageal adenocarcinoma: time for a new syn- surveillance. Clin Cancer Res 2007; 13: 659-665
thesis. Nat Rev Cancer 2010; 10: 87-101 38 Schulmann K, Sterian A, Berki A, Yin J, Sato F, Xu Y, Olaru A,
20 Shaheen NJ, Crosby MA, Bozymski EM, Sandler RS. Is Wang S, Mori Y, Deacu E, Hamilton J, Kan T, Krasna MJ, Beer
there publication bias in the reporting of cancer risk in Bar- DG, Pepe MS, Abraham JM, Feng Z, Schmiegel W, Green-
rett’s esophagus? Gastroenterology 2000; 119: 333-338 wald BD, Meltzer SJ. Inactivation of p16, RUNX3, and HPP1
21 Somerville M, Garside R, Pitt M, Stein K. Surveillance of occurs early in Barrett’s-associated neoplastic progression
Barrett’s oesophagus: is it worthwhile? Eur J Cancer 2008; 44: and predicts progression risk. Oncogene 2005; 24: 4138-4148
588-599 39 Jin Z, Cheng Y, Gu W, Zheng Y, Sato F, Mori Y, Olaru AV,
22 Manne U, Srivastava RG, Srivastava S. Recent advances in Paun BC, Yang J, Kan T, Ito T, Hamilton JP, Selaru FM,
biomarkers for cancer diagnosis and treatment. Drug Discov Agarwal R, David S, Abraham JM, Wolfsen HC, Wallace
Today 2005; 10: 965-976 MB, Shaheen NJ, Washington K, Wang J, Canto MI, Bhat-
23 Pepe MS, Etzioni R, Feng Z, Potter JD, Thompson ML, tacharyya A, Nelson MA, Wagner PD, Romero Y, Wang KK,

WJG|www.wjgnet.com 5678 December 7, 2010|Volume 16|Issue 45|

 Ong CAJ et al . Biomarkers in Barrett’s esophagus

Feng Z, Sampliner RE, Meltzer SJ. A multicenter, double- 55 Hong MK, Laskin WB, Herman BE, Johnston MH, Vargo
blinded validation study of methylation biomarkers for pro- JJ, Steinberg SM, Allegra CJ, Johnston PG. Expansion of the
gression prediction in Barrett’s esophagus. Cancer Res 2009; Ki-67 proliferative compartment correlates with degree of
69: 4112-4115 dysplasia in Barrett’s esophagus. Cancer 1995; 75: 423-429
40 Schlemper RJ, Riddell RH, Kato Y, Borchard F, Cooper 56 Going JJ, Keith WN, Neilson L, Stoeber K, Stuart RC, Wil-
HS, Dawsey SM, Dixon MF, Fenoglio-Preiser CM, Fléjou liams GH. Aberrant expression of minichromosome main-
JF, Geboes K, Hattori T, Hirota T, Itabashi M, Iwafuchi M, tenance proteins 2 and 5, and Ki-67 in dysplastic squamous
Iwashita A, Kim YI, Kirchner T, Klimpfinger M, Koike M, oesophageal epithelium and Barrett’s mucosa. Gut 2002; 50:
Lauwers GY, Lewin KJ, Oberhuber G, Offner F, Price AB, 373-377
Rubio CA, Shimizu M, Shimoda T, Sipponen P, Solcia E, 57 Trudgill NJ, Suvarna SK, Royds JA, Riley SA. Cell cycle
Stolte M, Watanabe H, Yamabe H. The Vienna classification regulation in patients with intestinal metaplasia at the gas-
of gastrointestinal epithelial neoplasia. Gut 2000; 47: 251-255 tro-oesophageal junction. Mol Pathol 2003; 56: 313-317
41 Pech O, Vieth M, Schmitz D, Gossner L, May A, Seitz G, 58 Bani-Hani K, Martin IG, Hardie LJ, Mapstone N, Briggs JA,
Stolte M, Ell C. Conclusions from the histological diagnosis Forman D, Wild CP. Prospective study of cyclin D1 overex-
of low-grade intraepithelial neoplasia in Barrett’s oesopha- pression in Barrett’s esophagus: association with increased
gus. Scand J Gastroenterol 2007; 42: 682-688 risk of adenocarcinoma. J Natl Cancer Inst 2000; 92: 1316-1321
42 Jankowski JA, Odze RD. Biomarkers in gastroenterology: 59 Eads CA, Lord RV, Kurumboor SK, Wickramasinghe K,
between hope and hype comes histopathology. Am J Gastro- Skinner ML, Long TI, Peters JH, DeMeester TR, Danenberg
enterol 2009; 104: 1093-1096 KD, Danenberg PV, Laird PW, Skinner KA. Fields of aber-
43 Rabinovitch PS, Longton G, Blount PL, Levine DS, Reid BJ. rant CpG island hypermethylation in Barrett’s esopha-
Predictors of progression in Barrett’s esophagus III: base- gus and associated adenocarcinoma. Cancer Res 2000; 60:
line flow cytometric variables. Am J Gastroenterol 2001; 96: 5021-5026
3071-3083 60 Jin Z, Cheng Y, Gu W, Zheng Y, Sato F, Mori Y, Olaru AV,
44 Galipeau PC, Cowan DS, Sanchez CA, Barrett MT, Emond Paun BC, Yang J, Kan T, Ito T, Hamilton JP, Selaru FM,
MJ, Levine DS, Rabinovitch PS, Reid BJ. 17p (p53) allelic Agarwal R, David S, Abraham JM, Wolfsen HC, Wallace
losses, 4N (G2/tetraploid) populations, and progression to MB, Shaheen NJ, Washington K, Wang J, Canto MI, Bhat-
aneuploidy in Barrett’s esophagus. Proc Natl Acad Sci USA tacharyya A, Nelson MA, Wagner PD, Romero Y, Wang KK,
1996; 93: 7081-7084 Feng Z, Sampliner RE, Meltzer SJ. A multicenter, double-
45 Fang M, Lew E, Klein M, Sebo T, Su Y, Goyal R. DNA abnor- blinded validation study of methylation biomarkers for pro-
malities as marker of risk for progression of Barrett’s esopha- gression prediction in Barrett’s esophagus. Cancer Res 2009;
gus to adenocarcinoma: image cytometric DNA analysis in 69: 4112-4115
formalin-fixed tissues. Am J Gastroenterol 2004; 99: 1887-1894 61 Wang JS, Guo M, Montgomery EA, Thompson RE, Cosby
46 Wong DJ, Paulson TG, Prevo LJ, Galipeau PC, Longton G, H, Hicks L, Wang S, Herman JG, Canto MI. DNA promoter
Blount PL, Reid BJ. p16(INK4a) lesions are common, early hypermethylation of p16 and APC predicts neoplastic pro-
abnormalities that undergo clonal expansion in Barrett’s gression in Barrett’s esophagus. Am J Gastroenterol 2009; 104:
metaplastic epithelium. Cancer Res 2001; 61: 8284-8289 2153-2160
47 Leedham SJ, Preston SL, McDonald SA, Elia G, Bhandari 62 Eads CA, Lord RV, Wickramasinghe K, Long TI, Kurum-
P, Poller D, Harrison R, Novelli MR, Jankowski JA, Wright boor SK, Bernstein L, Peters JH, DeMeester SR, DeMeester
NA. Individual crypt genetic heterogeneity and the origin TR, Skinner KA, Laird PW. Epigenetic patterns in the pro-
of metaplastic glandular epithelium in human Barrett’s oe- gression of esophageal adenocarcinoma. Cancer Res 2001; 61:
sophagus. Gut 2008; 57: 1041-1048 3410-3418
48 Reid BJ, Prevo LJ, Galipeau PC, Sanchez CA, Longton G, 63 Clark SJ, Harrison J, Paul CL, Frommer M. High sensitivity
Levine DS, Blount PL, Rabinovitch PS. Predictors of pro- mapping of methylated cytosines. Nucleic Acids Res 1994; 22:
gression in Barrett’s esophagus II: baseline 17p (p53) loss of 2990-2997
heterozygosity identifies a patient subset at increased risk for 64 Weber M, Davies JJ, Wittig D, Oakeley EJ, Haase M, Lam
neoplastic progression. Am J Gastroenterol 2001; 96: 2839-2848 WL, Schübeler D. Chromosome-wide and promoter-specific
49 Vogt N, Schönegg R, Gschossmann JM, Borovicka J. Benefit analyses identify sites of differential DNA methylation in
of baseline cytometry for surveillance of patients with Bar- normal and transformed human cells. Nat Genet 2005; 37:
rett’s esophagus. Surg Endosc 2010; 24: 1144-1150 853-862
50 Huang Q, Yu C, Klein M, Fang J, Goyal RK. DNA index 65 Huang TH, Laux DE, Hamlin BC, Tran P, Tran H, Lubahn
determination with Automated Cellular Imaging System DB. Identification of DNA methylation markers for human
(ACIS) in Barrett’s esophagus: comparison with CAS 200. breast carcinomas using the methylation-sensitive restric-
BMC Clin Pathol 2005; 5: 7 tion fingerprinting technique. Cancer Res 1997; 57: 1030-1034
51 Yu C, Zhang X, Huang Q, Klein M, Goyal RK. High-fidelity 66 Khulan B, Thompson RF, Ye K, Fazzari MJ, Suzuki M, Stasiek
DNA histograms in neoplastic progression in Barrett’s E, Figueroa ME, Glass JL, Chen Q, Montagna C, Hatchwell E,
esophagus. Lab Invest 2007; 87: 466-472 Selzer RR, Richmond TA, Green RD, Melnick A, Greally JM.
52 Li X, Galipeau PC, Sanchez CA, Blount PL, Maley CC, Comparative isoschizomer profiling of cytosine methylation:
Arnaudo J, Peiffer DA, Pokholok D, Gunderson KL, Reid the HELP assay. Genome Res 2006; 16: 1046-1055
BJ. Single nucleotide polymorphism-based genome-wide 67 Frigola J, Ribas M, Risques RA, Peinado MA. Methylome
chromosome copy change, loss of heterozygosity, and aneu- profiling of cancer cells by amplification of inter-methylated
ploidy in Barrett’s esophagus neoplastic progression. Cancer sites (AIMS). Nucleic Acids Res 2002; 30: e28
Prev Res (Phila Pa) 2008; 1: 413-423 68 Ushijima T, Morimura K, Hosoya Y, Okonogi H, Tatematsu
53 Paulson TG, Galipeau PC, Reid BJ. Loss of heterozygosity M, Sugimura T, Nagao M. Establishment of methylation-
analysis using whole genome amplification, cell sorting, and sensitive-representational difference analysis and isolation
fluorescence-based PCR. Genome Res 1999; 9: 482-491 of hypo- and hypermethylated genomic fragments in mouse
54 Wongsurawat VJ, Finley JC, Galipeau PC, Sanchez CA, liver tumors. Proc Natl Acad Sci USA 1997; 94: 2284-2289
Maley CC, Li X, Blount PL, Odze RD, Rabinovitch PS, Reid 69 Hatada I, Hayashizaki Y, Hirotsune S, Komatsubara H,
BJ. Genetic mechanisms of TP53 loss of heterozygosity in Mukai T. A genomic scanning method for higher organisms
Barrett’s esophagus: implications for biomarker validation. using restriction sites as landmarks. Proc Natl Acad Sci USA
Cancer Epidemiol Biomarkers Prev 2006; 15: 509-516 1991; 88: 9523-9527

WJG|www.wjgnet.com 5679 December 7, 2010|Volume 16|Issue 45|

 Ong CAJ et al . Biomarkers in Barrett’s esophagus

70 Bibikova M, Lin Z, Zhou L, Chudin E, Garcia EW, Wu B, son DH, Kelsen DP, Harpole D, Moore MB, Danenberg KD,
Doucet D, Thomas NJ, Wang Y, Vollmer E, Goldmann T, Danenberg PV, Meltzer SJ. Hypermethylated APC DNA in
Seifart C, Jiang W, Barker DL, Chee MS, Floros J, Fan JB. plasma and prognosis of patients with esophageal adeno-
High-throughput DNA methylation profiling using univer- carcinoma. J Natl Cancer Inst 2000; 92: 1805-1811
sal bead arrays. Genome Res 2006; 16: 383-393 88 Heeren PA, Kloppenberg FW, Hollema H, Mulder NH, Nap
71 Ibanez de Caceres I, Dulaimi E, Hoffman AM, Al-Saleem T, RE, Plukker JT. Predictive effect of p53 and p21 alteration
Uzzo RG, Cairns P. Identification of novel target genes by on chemotherapy response and survival in locally advanced
an epigenetic reactivation screen of renal cancer. Cancer Res adenocarcinoma of the esophagus. Anticancer Res 2004; 24:
2006; 66: 5021-5028 2579-2583
72 Laird PW. Principles and challenges of genome-wide DNA 89 Nakashima S, Natsugoe S, Matsumoto M, Kijima F, Take-
methylation analysis. Nat Rev Genet 2010; 11: 191-203 bayashi Y, Okumura H, Shimada M, Nakano S, Kusano C,
73 CancerStats. Cancer Research UK, 2009. Accessed on Baba M, Takao S, Aikou T. Expression of p53 and p21 is
25/11/2009. Available from: URL: http://info.cancerre- useful for the prediction of preoperative chemotherapeutic
searchuk.org/cancerstats/ effects in esophageal carcinoma. Anticancer Res 2000; 20:
74 Edge SB, Compton CC, Fritz AG, Greene FL, Trotti A. AJCC 1933-1937
Cancer Staging Manual. New York, NY: Springer, 2010 90 Raouf AA, Evoy DA, Carton E, Mulligan E, Griffin MM,
75 Hanahan D, Weinberg RA. The hallmarks of cancer. Cell Reynolds JV. Loss of Bcl-2 expression in Barrett’s dysplasia
2000; 100: 57-70 and adenocarcinoma is associated with tumor progression
76 Lagarde SM, ten Kate FJ, Richel DJ, Offerhaus GJ, van Lan- and worse survival but not with response to neoadjuvant
schot JJ. Molecular prognostic factors in adenocarcinoma chemoradiation. Dis Esophagus 2003; 16: 17-23
of the esophagus and gastroesophageal junction. Ann Surg 91 Bhandari P, Bateman AC, Mehta RL, Stacey BS, Johnson P,
Oncol 2007; 14: 977-991 Cree IA, Di Nicolantonio F, Patel P. Prognostic significance
77 Izzo JG, Wu TT, Wu X, Ensor J, Luthra R, Pan J, Correa A, of cyclooxygenase-2 (COX-2) expression in patients with
Swisher SG, Chao CK, Hittelman WN, Ajani JA. Cyclin D1 surgically resectable adenocarcinoma of the oesophagus.
guanine/adenine 870 polymorphism with altered protein BMC Cancer 2006; 6: 134
expression is associated with genomic instability and ag- 92 France M, Drew PA, Dodd T, Watson DI. Cyclo-oxygen-
gressive clinical biology of esophageal adenocarcinoma. J ase-2 expression in esophageal adenocarcinoma as a deter-
Clin Oncol 2007; 25: 698-707 minant of clinical outcome following esophagectomy. Dis
78 Wang KL, Wu TT, Choi IS, Wang H, Resetkova E, Correa Esophagus 2004; 17: 136-140
AM, Hofstetter WL, Swisher SG, Ajani JA, Rashid A, Albar- 93 Buskens CJ, Van Rees BP, Sivula A, Reitsma JB, Haglund C,
racin CT. Expression of epidermal growth factor receptor in Bosma PJ, Offerhaus GJ, Van Lanschot JJ, Ristimäki A. Prog-
esophageal and esophagogastric junction adenocarcinomas: nostic significance of elevated cyclooxygenase 2 expression
association with poor outcome. Cancer 2007; 109: 658-667 in patients with adenocarcinoma of the esophagus. Gastro-
79 Langer R, Von Rahden BH, Nahrig J, Von Weyhern C, Re- enterology 2002; 122: 1800-1807
iter R, Feith M, Stein HJ, Siewert JR, Höfler H, Sarbia M. 94 Izzo JG, Malhotra U, Wu TT, Ensor J, Luthra R, Lee JH,
Prognostic significance of expression patterns of c-erbB-2, Swisher SG, Liao Z, Chao KS, Hittelman WN, Aggarwal
p53, p16INK4A, p27KIP1, cyclin D1 and epidermal growth BB, Ajani JA. Association of activated transcription factor
factor receptor in oesophageal adenocarcinoma: a tissue mi- nuclear factor kappab with chemoradiation resistance and
croarray study. J Clin Pathol 2006; 59: 631-634 poor outcome in esophageal carcinoma. J Clin Oncol 2006;
80 Falkenback D, Nilbert M, Oberg S, Johansson J. Prognostic 24: 748-754
value of cell adhesion in esophageal adenocarcinomas. Dis 95 Gertler R, Doll D, Maak M, Feith M, Rosenberg R. Telomere
Esophagus 2008; 21: 97-102 length and telomerase subunits as diagnostic and prognostic
81 Brien TP, Odze RD, Sheehan CE, McKenna BJ, Ross JS. biomarkers in Barrett carcinoma. Cancer 2008; 112: 2173-2180
HER-2/neu gene amplification by FISH predicts poor sur- 96 Saad RS, El-Gohary Y, Memari E, Liu YL, Silverman JF.
vival in Barrett’s esophagus-associated adenocarcinoma. Endoglin (CD105) and vascular endothelial growth factor
Hum Pathol 2000; 31: 35-39 as prognostic markers in esophageal adenocarcinoma. Hum
82 Ross JS, McKenna BJ. The HER-2/neu oncogene in tumors Pathol 2005; 36: 955-961
of the gastrointestinal tract. Cancer Invest 2001; 19: 554-568 97 Nekarda H, Schlegel P, Schmitt M, Stark M, Mueller JD,
83 Aloia TA, Harpole DH Jr, Reed CE, Allegra C, Moore MB, Fink U, Siewert JR. Strong prognostic impact of tumor-asso-
Herndon JE 2nd, D’Amico TA. Tumor marker expression ciated urokinase-type plasminogen activator in completely
is predictive of survival in patients with esophageal cancer. resected adenocarcinoma of the esophagus. Clin Cancer Res
Ann Thorac Surg 2001; 72: 859-866 1998; 4: 1755-1763
84 D’Errico A, Barozzi C, Fiorentino M, Carella R, Di Simone 98 Darnton SJ, Hardie LJ, Muc RS, Wild CP, Casson AG. Tis-
M, Ferruzzi L, Mattioli S, Grigioni WF. Role and new per- sue inhibitor of metalloproteinase-3 (TIMP-3) gene is meth-
spectives of transforming growth factor-alpha (TGF-alpha) ylated in the development of esophageal adenocarcinoma:
in adenocarcinoma of the gastro-oesophageal junction. Br J loss of expression correlates with poor prognosis. Int J Can-
Cancer 2000; 82: 865-870 cer 2005; 115: 351-358
85 von Rahden BH, Stein HJ, Feith M, Pühringer F, Theisen 99 Brock MV, Gou M, Akiyama Y, Muller A, Wu TT, Mont-
J, Siewert JR, Sarbia M. Overexpression of TGF-beta1 in gomery E, Deasel M, Germonpré P, Rubinson L, Heitmiller
esophageal (Barrett’s) adenocarcinoma is associated with RF, Yang SC, Forastiere AA, Baylin SB, Herman JG. Prog-
advanced stage of disease and poor prognosis. Mol Carcinog nostic importance of promoter hypermethylation of mul-
2006; 45: 786-794 tiple genes in esophageal adenocarcinoma. Clin Cancer Res
86 Fukuchi M, Miyazaki T, Fukai Y, Nakajima M, Sohda M, 2003; 9: 2912-2919
Masuda N, Manda R, Tsukada K, Kato H, Kuwano H. 100 Baumann S, Keller G, Pühringer F, Napieralski R, Feith M,
Plasma level of transforming growth factor beta1 measured Langer R, Höfler H, Stein HJ, Sarbia M. The prognostic im-
from the azygos vein predicts prognosis in patients with pact of O6-Methylguanine-DNA Methyltransferase (MGMT)
esophageal cancer. Clin Cancer Res 2004; 10: 2738-2741 promotor hypermethylation in esophageal adenocarcinoma.
87 Kawakami K, Brabender J, Lord RV, Groshen S, Greenwald Int J Cancer 2006; 119: 264-268
BD, Krasna MJ, Yin J, Fleisher AS, Abraham JM, Beer DG, 101 van’t Veer LJ, Dai H, van de Vijver MJ, He YD, Hart AA,
Sidransky D, Huss HT, Demeester TR, Eads C, Laird PW, Il- Mao M, Peterse HL, van der Kooy K, Marton MJ, Witteveen

WJG|www.wjgnet.com 5680 December 7, 2010|Volume 16|Issue 45|

 Ong CAJ et al . Biomarkers in Barrett’s esophagus

AT, Schreiber GJ, Kerkhoven RM, Roberts C, Linsley PS, huijs-Pederson LA, Ten Kate FJ, Reitsma PH, van Kampen
Bernards R, Friend SH. Gene expression profiling predicts AH, Zwinderman AH, Baas F, van Lanschot JJ. Analysis of
clinical outcome of breast cancer. Nature 2002; 415: 530-536 gene expression identifies differentially expressed genes
102 Fan C, Oh DS, Wessels L, Weigelt B, Nuyten DS, Nobel AB, and pathways associated with lymphatic dissemination in
van’t Veer LJ, Perou CM. Concordance among gene-expres- patients with adenocarcinoma of the esophagus. Ann Surg
sion-based predictors for breast cancer. N Engl J Med 2006; Oncol 2008; 15: 3459-3470
355: 560-569 115 Luthra MG, Ajani JA, Izzo J, Ensor J, Wu TT, Rashid A,
103 Karlsson E, Delle U, Danielsson A, Olsson B, Abel F, Karls- Zhang L, Phan A, Fukami N, Luthra R. Decreased expres-
son P, Helou K. Gene expression variation to predict 10-year sion of gene cluster at chromosome 1q21 defines molecular
survival in lymph-node-negative breast cancer. BMC Cancer subgroups of chemoradiotherapy response in esophageal
2008; 8: 254 cancers. Clin Cancer Res 2007; 13: 912-919
104 Selaru FM, Zou T, Xu Y, Shustova V, Yin J, Mori Y, Sato F, 116 Luthra R, Wu TT, Luthra MG, Izzo J, Lopez-Alvarez E,
Wang S, Olaru A, Shibata D, Greenwald BD, Krasna MJ, Zhang L, Bailey J, Lee JH, Bresalier R, Rashid A, Swisher
Abraham JM, Meltzer SJ. Global gene expression profiling in SG, Ajani JA. Gene expression profiling of localized esopha-
Barrett’s esophagus and esophageal cancer: a comparative geal carcinomas: association with pathologic response to
analysis using cDNA microarrays. Oncogene 2002; 21: 475-478 preoperative chemoradiation. J Clin Oncol 2006; 24: 259-267
105 Greenawalt DM, Duong C, Smyth GK, Ciavarella ML, 117 Schauer M, Janssen KP, Rimkus C, Raggi M, Feith M,
Thompson NJ, Tiang T, Murray WK, Thomas RJ, Phillips Friess H, Theisen J. Microarray-based response prediction
WA. Gene expression profiling of esophageal cancer: compar- in esophageal adenocarcinoma. Clin Cancer Res 2010; 16:
ative analysis of Barrett's esophagus, adenocarcinoma, and 330-337
squamous cell carcinoma. Int J Cancer 2007; 120: 1914-1921 118 Duong C, Greenawalt DM, Kowalczyk A, Ciavarella ML,
106 Ntzani EE, Ioannidis JP. Predictive ability of DNA microar- Raskutti G, Murray WK, Phillips WA, Thomas RJ. Pre-
rays for cancer outcomes and correlates: an empirical as- treatment gene expression profiles can be used to predict
sessment. Lancet 2003; 362: 1439-1444 response to neoadjuvant chemoradiotherapy in esophageal
107 Michiels S, Koscielny S, Hill C. Prediction of cancer out- cancer. Ann Surg Oncol 2007; 14: 3602-3609
come with microarrays: a multiple random validation strat- 119 Langer R, Ott K, Specht K, Becker K, Lordick F, Burian M,
egy. Lancet 2005; 365: 488-492 Herrmann K, Schrattenholz A, Cahill MA, Schwaiger M,
108 Abu-Hanna A, Lucas PJ. Prognostic models in medicine. AI Hofler H, Wester HJ. Protein expression profiling in esopha-
and statistical approaches. Methods Inf Med 2001; 40: 1-5 geal adenocarcinoma patients indicates association of heat-
109 Royston P, Moons KG, Altman DG, Vergouwe Y. Prognosis shock protein 27 expression and chemotherapy response.
and prognostic research: Developing a prognostic model. Clin Cancer Res 2008; 14: 8279-8287
BMJ 2009; 338: b604 120 Wu X, Gu J, Wu TT, Swisher SG, Liao Z, Correa AM, Liu J,
110 Moons KG, Royston P, Vergouwe Y, Grobbee DE, Altman Etzel CJ, Amos CI, Huang M, Chiang SS, Milas L, Hittelman
DG. Prognosis and prognostic research: what, why, and WN, Ajani JA. Genetic variations in radiation and chemo-
how? BMJ 2009; 338: b375 therapy drug action pathways predict clinical outcomes in
111 Moons KG, Altman DG, Vergouwe Y, Royston P. Prognosis esophageal cancer. J Clin Oncol 2006; 24: 3789-3798
and prognostic research: application and impact of prognos- 121 Mathé EA, Nguyen GH, Bowman ED, Zhao Y, Budhu A,
tic models in clinical practice. BMJ 2009; 338: b606 Schetter AJ, Braun R, Reimers M, Kumamoto K, Hughes D,
112 Altman DG, Vergouwe Y, Royston P, Moons KG. Prognosis Altorki NK, Casson AG, Liu CG, Wang XW, Yanaihara N,
and prognostic research: validating a prognostic model. Hagiwara N, Dannenberg AJ, Miyashita M, Croce CM, Har-
BMJ 2009; 338: b605 ris CC. MicroRNA expression in squamous cell carcinoma
113 Peters CJ, Rees JR, Hardwick RH, Hardwick JS, Vowler SL, and adenocarcinoma of the esophagus: associations with
Ong CA, Zhang C, Save V, O’Donovan M, Rassl D, Alder- survival. Clin Cancer Res 2009; 15: 6192-6200
son D, Caldas C, Fitzgerald RC. A 4-Gene Signature Predicts 122 Pasello G, Agata S, Bonaldi L, Corradin A, Montagna M,
Survival of Patients With Resected Adenocarcinoma of the Zamarchi R, Parenti A, Cagol M, Zaninotto G, Ruol A, An-
Esophagus, Junction, and Gastric Cardia. Gastroenterology cona E, Amadori A, Saggioro D. DNA copy number altera-
2010; Epub ahead of print tions correlate with survival of esophageal adenocarcinoma
114 Lagarde SM, Ver Loren van Themaat PE, Moerland PD, Gil- patients. Mod Pathol 2009; 22: 58-65

S- Editor Wang JL L- Editor Kerr C E- Editor Ma WH

WJG|www.wjgnet.com 5681 December 7, 2010|Volume 16|Issue 45|